mlsanityData to 7 Oct 2026

MoE for time series and graphs

114 papers0 this weekMedium confidence in this grouping: mostly one topic, with blurry edgesMedium confidence in this grouping: mostly one topic, with blurry edges · About

Keywords: forecasting, series, time series, time, graph

Key facts. MoE for time series and graphs: 114 papers on arXiv since 2019 (data through 7 Oct 2026). 18 of its 114 mapped papers (16%) cite “Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts” (2024, arXiv 2409.16040). 12 of its 114 mapped papers (11%) cite “Moirai-MoE: Empowering Time Series Foundation Models with Sparse Mixture of Experts” (2024, arXiv 2410.10469). 2 new papers 17 Sep–7 Oct.

What’s new

17 Sep – 7 Oct 2026

2 new since 17 Sep

  1. Raw-Routed Mixture of Adapters: A Causal Intervention for Routing Collapse in Time Series Foundation Models
  2. OceanMoE: Structured Conditional Sparse Computation for Long-Horizon Multivariate Ocean Forecasting

Foundations

Most cited by papers here

  1. Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts2024cited by 18 here
  2. Moirai-MoE: Empowering Time Series Foundation Models with Sparse Mixture of Experts2024cited by 12 here
  3. Graph Mixture of Experts: Learning on Large-Scale Graphs with Explicit Diversity Modeling2023cited by 6 here
  4. Spatial Mixture-of-Experts2022cited by 4 here
  5. TESTAM: A Time-Enhanced Spatio-Temporal Attention Model with Mixture of Experts2024cited by 3 here

Lineage

The papers this direction builds on, oldest first.

Each paper on the line cites the one before.

  1. 2017Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layercited by 35 here
  2. 2020GShard: Scaling Giant Models with Conditional Computation and Automatic Shardingcited by 16 here
  3. 2021Efficient Large Scale Language Modeling with Mixtures of Expertscited by 3 here
  4. 2022DeepSpeed-MoE: Advancing Mixture-of-Experts Inference and Training to Power Next-Generation AI Scalecited by 3 here
  5. 2024DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Modelscited by 7 here
  6. 2024Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Expertscited by 18 here
18 more it builds on, not in this chain
  1. 2017Hard Mixtures of Experts for Large Scale Weakly Supervised Visioncited by 2 here
  2. 2018Deep Mixture of Experts via Shallow Embeddingcited by 3 here
  3. 2021Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsitycited by 29 here
  4. 2021Scaling Vision with Sparse Mixture of Expertscited by 12 here
  5. 2021GLaM: Efficient Scaling of Language Models with Mixture-of-Expertscited by 11 here
  6. 2022Mixture-of-Experts with Expert Choice Routingcited by 12 here
  7. 2022On the Representation Collapse of Sparse Mixture of Expertscited by 2 here
  8. 2022Multimodal Contrastive Learning with LIMoE: the Language-Image Mixture of Expertscited by 4 here
  9. 2022Sparse Mixture-of-Experts are Domain Generalizable Learnerscited by 2 here
  10. 2022Spatial Mixture-of-Expertscited by 4 here
  11. 2023Graph Mixture of Experts: Learning on Large-Scale Graphs with Explicit Diversity Modelingcited by 6 here
  12. 2023From Sparse to Soft Mixtures of Expertscited by 4 here
  13. 2024Mixtral of Expertscited by 7 here
  14. 2024TESTAM: A Time-Enhanced Spatio-Temporal Attention Model with Mixture of Expertscited by 3 here
  15. 2024Boosting Continual Learning of Vision-Language Models via Mixture-of-Experts Adapterscited by 3 here
  16. 2024A Survey on Mixture of Experts in Large Language Modelscited by 3 here
  17. 2024Moirai-MoE: Empowering Time Series Foundation Models with Sparse Mixture of Expertscited by 12 here
  18. 2024DeepSeek-V3 Technical Reportcited by 4 here

The 24 most cited of 88.

Browse

30 of 114, ranked 6–35

  1. 6FreqMoE: Enhancing Time Series Forecasting through Frequency Decomposition Mixture of Experts20252
  2. 7Node-wise Filtering in Graph Neural Networks: A Mixture of Experts Approach20242
  3. 8Interpretable Cascading Mixture-of-Experts for Urban Traffic Congestion Prediction20242
  4. 9Graph Mixture of Experts and Memory-augmented Routers for Multivariate Time Series Anomaly Detection20242
  5. 10GraphMETRO: Mitigating Complex Graph Distribution Shifts via Mixture of Aligned Experts20232
  6. 11AME-TS: Anchored Mixture-of-Experts for Time Series Forecasting20261
  7. 12MoFE-Time: Mixture of Frequency Domain Experts for Time-Series Forecasting Models20251
  8. 13Wavelet Mixture of Experts for Time Series Forecasting20251
  9. 14GateTS: Versatile and Efficient Forecasting via Attention-Inspired routed Mixture-of-Experts20251
  10. 15Mixture of Link Predictors on Graphs20241
All 30
  1. 16EWMoE: An effective model for global weather forecasting with mixture-of-experts20241
  2. 17Continual Traffic Forecasting via Mixture of Experts20241
  3. 18A Time Series is Worth Five Experts: Heterogeneous Mixture of Experts for Traffic Flow Prediction20241
  4. 19ST-ExpertNet: A Deep Expert Framework for Traffic Prediction20221
  5. 20MixTTE: Multi-Level Mixture-of-Experts for Scalable and Adaptive Travel Time Estimation20260
  6. 21FaST: Efficient and Effective Long-Horizon Forecasting for Large-Scale Spatial-Temporal Graphs via Mixture-of-Experts20260
  7. 22Scalable Heterogeneous Graph Learning via Heterogeneous-aware Orthogonal Prototype Experts20260
  8. 23M2FMoE: Multi-Resolution Multi-View Frequency Mixture-of-Experts for Extreme-Adaptive Time Series Forecasting20260
  9. 24One-Shot Price Forecasting with Covariate-Guided Experts under Privacy Constraints20260
  10. 25Diff-MN: Diffusion Parameterized MoE-NCDE for Continuous Time Series Generation with Irregular Observations20260
  11. 26Multi-Modal Time Series Prediction via Mixture of Modulated Experts20260
  12. 27Seg-MoE: Multi-Resolution Segment-wise Mixture-of-Experts for Time Series Forecasting Transformers20260
  13. 28MoHETS: Long-term Time Series Forecasting with Mixture-of-Heterogeneous-Experts20260
  14. 29UrbanMoE: A Sparse Multi-Modal Mixture-of-Experts Framework for Multi-Task Urban Region Profiling20260
  15. 30SEER: Transformer-based Robust Time Series Forecasting via Automated Patch Enhancement and Replacement20260
  16. 31Let Experts Feel Uncertainty: A Multi-Expert Label Distribution Approach to Probabilistic Time Series Forecasting20260
  17. 32Generalizing GNNs with Tokenized Mixture of Experts20260
  18. 33TiMi: Empower Time Series Transformers with Multimodal Mixture of Experts20260
  19. 34Switch-Hurdle: A MoE Encoder with AR Hurdle Decoder for Intermittent Demand Forecasting20260
  20. 35Fed-GAME: Personalized Federated Learning with Graph Attention Mixture-of-Experts For Time-Series Forecasting20260